In our recent longitudinal CT/MRI studies of normal elderly and Alzheimers' disease (AD) patients using a special negative angle protocol, we found that hippocampal atrophy occurs very early in the natural history of AD and accurately predicts incipient dementia. Once AD becomes clinically evident, our longitudinal PET and CT research has consistently identified lateral temporal lobe degenerative changes as most salient. Moreover, in AD we have identified relationships between in vivo hippocampal atrophy, hypothalamic-pituitary-adrenal (HPA) axis dysregulation and PET glucose utilization deficits. Further, at postmortem, we find reduced brain microvessel glucose uptake. These studies have led us to propose a "brain glucose starvation hypothesis" in the pathogenesis of AD that is intimately related to hippocampal dysfunction. The proposed study will attempt to use PET to extend the CT and MRI finding that hippocampal change is the earliest brain change in AD. The main objective is to study individuals that are either cognitively normal or minimally impaired clinically (questionable dementia cases) using high resolution PET and MRI measures of the hippocampus, and comprehensive clinical and neuropsychological examinations. Each of the two groups will have 55 subjects selected and 50% from each group will have MRI evidence of hippocampal atrophy at baseline evaluation. For each participant, longitudinal follow-up will be conducted at 24 months, providing a baseline and one follow-up over the proposed 4-year study. The main hypothesis to be tested is that hippocampal changes will predict both the future development of an AD syndrome and the neocortical metabolic changes associated with AD. Early detection of AD will facilitate testing treatments under conditions of less brain damage.